acknowledgements : chris snyder/ncar
Post on 05-Feb-2016
Embed Size (px)
DESCRIPTIONImpact of Assimilating GPS RO refractivity on Tropical Cyclone Position and Intensity Analyses and Forecasts Liu Hui, Bill Kuo, and Jeff Anderson NCAR. Acknowledgements : Chris Snyder/NCAR. Outline. Brief overview of GPS RO soundings Hurricane Ernesto (2006): - PowerPoint PPT Presentation
Impact of Assimilating GPS RO refractivity on Tropical Cyclone Position and Intensity Analyses and Forecasts
Liu Hui, Bill Kuo, and Jeff Anderson
Acknowledgements: Chris Snyder/NCAR
OutlineBrief overview of GPS RO soundingsHurricane Ernesto (2006): Explore how RO refractivity impacts analyses of Q, T, and wind with EnKF and if EnKF can make meaningful multivariate covariance estimate in convective environment of tropicsAll typhoons of 2008: Examine impact of RO data on forecasts of the typhoonsSummary
GPS Radio Occultation (GPS RO)Basic measurement principle: Deduce atmospheric refractivity properties based on measurement of radio wave phase/time delay and amplitude.
Observed Atmospheric Volume L~300 kmZ~1 kmRO refractivity data:1. High vertical and coarse horizontal resolution.2. Refractivity can be retrieved from the signals time delay.3. Determined by water vapor and temperature in the troposphere.
Challenges to Assimilation of GPS ROfor TC ForecastsGPS RO data available for clear and cloudy skies.Mainly water vapor information in the lower troposphere.Complicated RO refractivity forecast error variances and correlations with T and winds.Hard to describe the complicated covariance in a static covariance function.Ensemble forecasts provide an alternative way to estimate the full multivariate and flow-dependent error covariance.
Assimilation of RO Refractivity for TC Forecasts with Ensemble DA at NCARMultiple TC cases studied: Hurricane Ernesto (2006) Typhoon Shanshan (2007) All typhoons of 2008: Nuri, Sinlaku, Hagupit, and Jangmi Typhoon Morakot (2009)
Results:Assimilation of RO refractivity in cycling mode produces meaningful corrections to large-scale environment water vapor and winds around TCs.RO data in the lower troposphere has positive impact.RO data reduces TC track and intensity forecast errors.
Analyses for Hurricane Ernesto (2006, August 25-30)Analysis 2-hourly, cycling August 21-26, 2006.GPSonly run: Assimilate ONLY RO refractivity data.GPSonly>6km run: Assimilate ONLY RO refractivity data higher than 6km.FCST run: Ensemble forecasts from NCEP FNL starting at 00Z August 21 with no additional assimilation.Radiosondes/dropsondes held to validate the analysesWRF-ARW/DART ensemble assimilation system.32 ensembles, 36km grid over Atlantic, 35 levels.Initial and boundary conditions from NCEP FNL global analyses; ensemble is created by WRF/3DVAR covariance.
Analysis of 850 hPa Wind and Total Column Cloud Water(06Z August 23, 36 hours before Ernestos genesis, GPSonly run)Convergence and convection in Ernestos genesis area.
Daily Analysis Increments of Q and T (GPSonly run)(850 hPa, August 23, 2 days before Ernestos genesis)
Daily Analysis Increments of Wind(August 23, 2 days before Ernestos genesis, m/s)
GPSonly>6km run700 hPaGPSonly run250 hPaStrong correlation between GPS in the lower troposphere and winds at all levels.
Forecast Difference of GPSonly-FCST (700 hPa)
00Z 24 AugustWater vapor 06Z 23 August Wind00Z 25 August Ernestos genesis time
Wind Forecast Differences (250 hPa)
GPSonly-FCST 06Z 23 August GPSonly>6km-FCST00Z 24 August00Z 25 August Ernestos genesis time
2-hour Forecast RMS fit to Radiosondes(Averaged over tropical Atlantic, 21-26 August, 2006)
48-hour Forecasts of Ernesto (2006) with Assimilation of GPS and Conventional ObservationsWith GPSWithout GPSActual storm24h FCST48h FCST
48-hour Forecasts of Ernesto (2006) with Assimilation of GPS and Conventional ObservationsWith GPSWith GPS>6kmActual storm24h FCST48h FCSTAssimilation of GPS > 6km shows less positive impact
Distribution of RO Data (September 8, 2008)
96 RO profiles, 50% penetrated into the lowest 1km layer
Assimilation Experiments for 2008 TyphoonsFour major typhoons occurred in 2008 over W Pacific.Analyses cycling 11 August 30 September 2008; 3-hourly.CTRL run: Assimilate conventional observations.GPS run: Assimilate conventional data + RO refractivity.Conventional observations from NCEP GFS: rawinsonde T, Q and wind, cloud winds, aircraft T and Q, surface pressure, TC center positions.NO TC bogus and radiance data used.WRF-ARW/DART ensemble assimilation system.96 ensembles, 27km grid over W. Pacific (95E- 165E, 0-50N).Initial and boundary conditions from NCEP global analyses; initial ensemble created by WRF/3DVAR covariance.48 hour ensemble mean forecasts from 00Z and 12Z daily.
Ensemble Mean 48H Forecasts of Intensity and Track Error(Valid at 00Z and 12Z, 8-26 September, 2008)
Average track error:CTRL: 323 km GPS: 289 km Average track error:CTRL: 366 km GPS: 339 km Typhoon Sinlaku (Sept 8-22) Typhoon Hagupit (Sept 16-26)
Ensemble Mean 48h Forecasts of Intensity and Track Error(Valid at 00Z and 12Z daily)
Average error:CTRL: 300 km GPS: 249 km Average error:CTRL: 254 km GPS: 231 km Typhoon Jangmi (Sept 24-30) Typhoon Nuri (Aug 17-23)
Analysis RMS Fits to Dropsondes/Radiosondes(domain averaged 11 August 30 September, 2008)
SummaryUse of GPS data improves forecasts of TC track and intensity. (In average, track errors are reduced ~11% for 48h forecasts)RO refractivity improves analysis of water vapor and wind in convective environment of TCs.Benefit of RO data is reduced when RO data in the lower troposphere is ignored.
****Top: Schematic depiction of tubular volume over which atmosphere contributes information to a single occultation phase (ray) measurement. The intensity of shading in the tube represents the relative weighting of atmospheric properties that contribute to the value retrieved at the center of the tube. For typical atmospheric structures, L and Z are approximately 300 and 1 km respectively.
Bottom: Typical along-track weighting function for a single radio occultation measurement (from Melbourne et al, 1994). Most of the information is contributed by a mesoscale atmospheric volume centered at the ray tangent point.
***There is a question or debate on if the GPS data in the lower troposphere is helpful because of the bias of the GPS data in the lowest 2km.However, model also has large error in the lower troposphere.*******The next three charts show the progression of Ernesto over the 102 hours of the NCAR forecast. Images on the left are actual satellite photos of the event. Images on the right were synthesized from the forecast made without GPSRO data, while the images in the center are from the forecast that assimilated the 15 GPSRO profiles. The elapsed time since the beginning of the forecast is shown on the left-hand images -- in this case, 6 and 30 hours.*******